[mesa-users] helium main sequence binary

[Josiah Schwab]

Hi Ying,

> 1. About inlist1, I use some parameters by default, and add one
> command line: mesa_dir = ‘ ‘. It seems to use mesa r-7624 environment.
> So can you please explain more about this command ?

The mesa_dir inlist variable lets you override the enviroment variable
MESA_DIR.  If MESA_DIR is set to the version of MESA that you want to
use, you don't need to 

> 2. How to make mesa run fast ? I though one way is to change the
> time_step, are there any other controls ?

As an rough heuristic, you can think about a MESA run taking

  (# of timesteps) x (# of grid points) x (# of isotopes)

This is extremely wrong in detail, but it does give the intuition that
two other ways to decrease the run time are to use smaller nuclear
networks and fewer grid points.  (There are many other things will
effect the run time, especially given that you are doing binaries.)

If you're not doing late-stage burning, you're probably already using
small networks and so there's not much to do there.

The zero-th order control for the number grid points is

 http://mesa.sourceforge.net/controls_defaults.html#mesh_delta_coeff

and there are many other related (more-specific) controls.

And no matter what you're working on, It is ESSENTIAL to confirm that
your results are independent of your choices for the spatial and
temporal resolution.  It is easy to get MESA to give you lots
and lots of numbers.  It is hard to make sure those numbers are
physically meaningful.

> l put the new inlists below if you want to check.  Your reply helped
> indeed, and thank you so much again.

Sorry, I don't have time to look at these new inlists.

Best,
Josiah